Media Flow Converter for Use in Real-Time Delivery Transactions

1. A computer-implemented system for coordinated media streaming comprising: a processor; a media flow converter for use with a server for simultaneously delivering media from multiple audio services to multiple Real Time Protocol (RTP) clients, the media flow converter including: a plurality of audio objects, each audio object having an audio queue for receiving media from one of the multiple audio services and an audio channel for directing a flow of media received from one of the multiple audio services to one of the multiple RTP clients, wherein each audio object has a non-real time thread for receiving media, building the audio queue, and packetizing received media into RTP packets to be placed on the audio queue; and a single RTP producer having a single real-time thread and a timer for coordinating a delivery of RTP packets from each of the audio objects to a corresponding RTP client based on a delivery schedule to achieve real-time delivery; wherein said RTP producer sleeps for a pre-specified interval, and upon wake, prioritizes service delivery based on said audio objects' wait time.

2. The system of claim 1 , wherein said delivery schedule complies with real-time requirements of said RTP client for providing continuous real-time delivery of media from said server to said RTP client.

3. The system of claim 1 , wherein the timer is a native timer that determines a sleep time and a wait time for each audio object.

4. The system of claim 3 , wherein said RTP producer updates said delivery schedule based on a wait time of each audio object, wherein said wait time is the amount of time an audio object has been waiting to send RTP packets.

5. The system of claim 1 , wherein said audio services can be one of a text-to-speech service, an audio processing service, and a music processing service.

6. The system of claim 1 , wherein said audio services are provided by a WebSphere Voice Server running on said server, wherein said server is a WebSphere Application Server, and wherein WebSphere Voice Server is integrated with said WebSphere Application Server for providing a mix of java transaction based processing and soft-real-time processing for interfacing with said media converter using a J2EE framework.

7. The system of claim 6 , wherein WebSphere Voice Server provides speech recognition and synthesis support to at least one said media converter hosting a Web-based VoiceXML application, wherein the support is for at least one real-time continuous media stream connecting said WebSphere Application Server with said Web-based VoiceXML application.

8. The system of claim 6 , wherein said single real-time thread controls communication between said WebSphere Voice Server and WebSphere Application Server for achieving real-time delivery.

9. A computer-implemented method for coordinated streaming, the method comprising the steps of: configuring the computer to provide a media flow converter for use with a server for simultaneously delivering media from multiple audio services to multiple Real Time Protocol (RTP) clients, the media flow converter including: a plurality of audio objects, each audio object having an audio queue for receiving media from one of the multiple audio services and an audio channel for directing a flow of media received from one of the multiple audio services to one of the multiple RTP clients, wherein each audio object has a non-real time thread for receiving media, building the audio queue, and packetizing the received media into RTP packets to be placed on the audio queue; and a single RTP producer having a single real-time thread and a timer; receiving media from at least one of the multiple audio services to at least one of the plurality of audio objects; each audio object packetizing received media into RTP packets and placing the RTP packets on the audio queue of the audio object; the RTP producer accessing the audio objects based on a delivery schedule; and the RTP producer coordinating a delivery of RTP packets from each of the audio objects to a corresponding RTP client in accordance with the delivery schedule to achieve real-time delivery; wherein said RTP producer sleeps for a pre-specified interval, and upon wake, prioritizes service delivery based on said wait time.

10. The method of claim 9 , wherein said delivery schedule complies with real-time requirements of each RTP client thereby providing continuous real-time service delivery.

11. The method of claim 9 , further comprising: determining at least one wait time that an audio object has been waiting to send RTP packets; and updating said delivery schedule in view of said wait time.

12. The method of claim 11 , wherein the timer is a native timer which determines a sleep time and wait time for an audio object.